Objective: To assess the relative risks for bleeding with thrombolytic therapy in patients who are managed using pulmonary angiograms compared with those managed using noninvasive tests, primarily the ventilation-perfusion lung scan. Design: A decision analysis based on data from other studies. Methods: The risk for major bleeding in patients with pulmonary embolism who receive thrombolytic therapy after a noninvasive diagnosis was assessed from complications of thrombolytic therapy in patients with myocardial infarction, assuming that the same risk ratio for major bleeding when comparing an invasive with a noninvasive approach applied to patients with pulmonary embolism. The risk ratio was 3.3 (95% CI, 1.5 to 9.8) for major bleeding in patients with myocardial infarction. One or more major complications of pulmonary angiography occurred in 1.3% of patients (CI, 0.6% to 1.9%). Results: The average reported risk was 14% (18 of 129 patients) (CI, 7.9% to 20.1%) for major bleeding in patients who had pulmonary angiography before receiving tissue plasminogen activator (tPA). The estimated risk was 4.2% (estimated CI, 1.4% to 9.3%) for major bleeding with tPA after a noninvasive diagnosis of pulmonary embolism. Assuming a risk of 1.3% for major complications from pulmonary angiography, a risk for major hemorrhage of 14.0% for an invasive diagnosis, and a risk of 4.2% for a noninvasive diagnosis, fewer complications would occur with noninvasive management if the prevalence of pulmonary embolism exceeded 21%. Conclusion: Among patients with suspected pulmonary embolism who are candidates for thrombolytic therapy, it is safer to use noninvasive diagnostic tests in many patients.